_345

Doesn't this necessitate installing a malicious python module, which is pretty unlikely unless a supply chain attack occurs right?

It would be really interesting to see if incest goes down after this using now as a baseline...

Seeing lots of content on the web about pangram's unreliability

Trying to be a little less negative than the guy that got flagged, I too don't understand the motivation behind these projects. I've seen a hundred of them at this point- and each of them is probably worse and has less learning value than the one Andrej Karpathy made to teach people the building blocks involved in a GPT

So, why do people keep making these? Asking genuinely

Are there days when you forget to apply your framework or you feel lazy and cheat/shortcut parts/all of the framework?

would you share a bit about the style of the code reviews and how its initiated? im trying to build something similar for my company

I love Eric. He is up next for sure.

I'm really skeptical you can ship ~10 PRs a day per person unless these PRs are tiny pieces of one feature or all of them are tiny bugs that are each a 3 line fix that you could review instantly. Otherwise how can you confirm that the AI even did the right fix or feature correctly? That you even wanted that feature done that way?

How does it compare performance wise? Feature parity? Can you spawn shells in the terminal pane or is there no terminal pane and you're expected to being your own multiplexer?

you get used to it almost instantly like within 1 hour of use. But I also have the OnePlus Open which has probably the best hidden crease you can buy besides Oppo which is just the company OnePlus came from

I've taken my OnePlus Open to the beach and it was fine. Multiple times. I don't use any case for my phone

Isn't it wasteful to have everyone go off and remake the same software with slightly different bugs in it instead of one that fixes almost all of the bugs

Yeah even if the software is good there's just something sad about realizing you're reading AI generated text. Even if its proofread and factually correct it just comes across as cheap and low effort, moreso when the content gets duplicated in a pointless FAQ that just restates everything you just read

Yes I filed a support ticket with them and explained that their system is broken and they just did not care. I explained how it was impossible for me to use it 3 times already as I've only made 2 chatgpt accounts EVER, and only recalling entering my phone number for one of the two chatgpt accounts. I told them that this issue locked me out of codex and chatgpt for work and they weren't willing to do anything about it. Totally useless support.

I ended up borrowing my gf's phone number just so I could get access for work. Ridiculous

Hope to be a part of this someday...

I'm pretty sure the lemon lime game referred to in the article is supposed to be some trick where it always says you're wrong and you have to cheat (maybe in browser developer tools) to get a 7/7

"87.3%

Share of the maximum achievable score our GRPO-trained 9B open-source model reached on catalog review, vs 76.9% for the best frontier configuration: a 13.5% relative improvement over the frontier, and 36% over its own untrained base (64.2%). The five frontier models, even with optimized prompts, plateaued within a tenth of a point of each other; the trained specialist cleared that ceiling."

_______

This seems very unlikely to be true. I have a lot of skepticism that frontier models like GPT 5.5 that are likely 2T+ parameters in size only got about 12% more accurate than an untrained 9b parameter LLM.

i got half way through the readme for this project and had the same thought and got sad and just left the page

The oneplus open (2023) is such a great phone, what a shame

What inspired you to make this?

i dont understand why you would use this. i think it needs more examples

I feel like its only useful if the work you are doing doesn't have correctness as a high priority. If your work is okay with something only mostly being correct and it can just be slopped together over time then yes throw 500 agents at it. But otherwise you can't really review all that work as a single human and will quickly run out of capacity

This makes so much sense as to why I've always felt that Opus 4.8 was leagues ahead of GPT 5.5. It's so good at taking underspecified requirements and filling in the gaps with sensible approaches for your project

I did the same thing but it's 25 feet hahaha, love to see this

Just you as an adult. It's always been that infantile

any guesses how?

I'm not letting Jenna from HR log into my personal machine with access to all of my lifelong data though. I do let my claude bypass permissions though

Best comment in this thread

In my case no, I actually saw worse performance with fable medium and switched back to opus high and xhigh

way worse things can happen than your machine being bricked, if a malicious actor can weaponize an agent to do their bidding